https://github.com/python/cpython/commit/46cb6340d7bad955edfc0a20f6a52dabc03... commit: 46cb6340d7bad955edfc0a20f6a52dabc03b0932 branch: main author: Alexander Shadchin <shadchin@yandex-team.com> committer: Eclips4 <kirill.bast9@mail.ru> date: 2025-01-03T18:47:58Z summary: gh-127903: Fix a crash on debug builds when calling `Objects/unicodeobject::_copy_characters`` (#127876) files: A Misc/NEWS.d/next/Core_and_Builtins/2024-12-13-14-17-24.gh-issue-127903.vemHSl.rst M Lib/test/test_str.py M Objects/unicodeobject.c diff --git a/Lib/test/test_str.py b/Lib/test/test_str.py index 4de6c1cba152bd..d1c9542c7d1317 100644 --- a/Lib/test/test_str.py +++ b/Lib/test/test_str.py @@ -7,6 +7,7 @@ """ import _string import codecs +import datetime import itertools import operator import pickle @@ -1908,6 +1909,12 @@ def test_utf8_decode_invalid_sequences(self): self.assertRaises(UnicodeDecodeError, (b'\xF4'+cb+b'\xBF\xBF').decode, 'utf-8') + def test_issue127903(self): + # gh-127903: ``_copy_characters`` crashes on DEBUG builds when + # there is nothing to copy. + d = datetime.datetime(2013, 11, 10, 14, 20, 59) + self.assertEqual(d.strftime('%z'), '') + def test_issue8271(self): # Issue #8271: during the decoding of an invalid UTF-8 byte sequence, # only the start byte and the continuation byte(s) are now considered diff --git a/Misc/NEWS.d/next/Core_and_Builtins/2024-12-13-14-17-24.gh-issue-127903.vemHSl.rst b/Misc/NEWS.d/next/Core_and_Builtins/2024-12-13-14-17-24.gh-issue-127903.vemHSl.rst new file mode 100644 index 00000000000000..ad479b52d1675c --- /dev/null +++ b/Misc/NEWS.d/next/Core_and_Builtins/2024-12-13-14-17-24.gh-issue-127903.vemHSl.rst @@ -0,0 +1,2 @@ +``Objects/unicodeobject.c``: fix a crash on DEBUG builds in ``_copy_characters`` +when there is nothing to copy. diff --git a/Objects/unicodeobject.c b/Objects/unicodeobject.c index 9f0a4d4785eda6..5e532ce0f348c4 100644 --- a/Objects/unicodeobject.c +++ b/Objects/unicodeobject.c @@ -1463,11 +1463,14 @@ _copy_characters(PyObject *to, Py_ssize_t to_start, assert(PyUnicode_Check(from)); assert(from_start + how_many <= PyUnicode_GET_LENGTH(from)); - assert(PyUnicode_Check(to)); - assert(to_start + how_many <= PyUnicode_GET_LENGTH(to)); + assert(to == NULL || PyUnicode_Check(to)); - if (how_many == 0) + if (how_many == 0) { return 0; + } + + assert(to != NULL); + assert(to_start + how_many <= PyUnicode_GET_LENGTH(to)); from_kind = PyUnicode_KIND(from); from_data = PyUnicode_DATA(from);